Re: [poe] isNeccessaryFor

Re @riannella comment before:

re 1: That they _must_ be checked at the time ... is not defined. And 
if they _need_ to be checked at the time ... depends on the 
constraint: if the constraints limits the use of an asset to the time 
after a specific date and time one can draw conclusions from that 
already hours before - e.g. putting the story in a publication queue 
with a publication time set to the time of the constraint.
And that's exactly the goal of my proposal: to define explicitly that 
this constraint must be check at the time of executing the action.

re 2: sorry, this definition does not cover the issue of having no 
result from the referred Constraint. Where is it defined that not 
being able to evaluate a constraint defaults to false - and what are 
the conditions for "not being able"?

re 3: The definition is ok, but it does not cover the case that the 
logical expression cannot be evaluated. Same as in 2.

re 4: ok, item 8 of the Constraint Relations processing model can be 
interpreted as defining an and-relationship.

re 5 and 6: that example shows a strange/wrong design of a constraint.
 My concerns are about not being able to process a reasonable 
constraint. And with my IPTC hat on I say that e.g. the NewsML-G2 
standard (for the exchange of news items) covers syntax errors and 
semantic errors relevant for the processing of a news item.

I can't see how this comment explains the need for isNecessaryFor or 
in general a rule, that one constraint must deliver a result before 
another constraint is processed - what use case cannot be solved 
without this rule? I see more need in indicating a constraint must be 
processed at the time of exercising an action, see my count example.


-- 
GitHub Notification of comment by nitmws
Please view or discuss this issue at 
https://github.com/w3c/poe/issues/98#issuecomment-276915655 using your
 GitHub account

Received on Thursday, 2 February 2017 10:04:04 UTC